New Method for Assimilable Organic Carbon Determination Using

Mar 31, 2005 - Citing Articles; Related Content ... Pyrosequencing Reveals Bacterial Communities in Unchlorinated Drinking Water Distribution System: ...
0 downloads 7 Views 191KB Size
Download PDF
Environ. Sci. Technol. 2005, 39, 3289-3294

New Method for Assimilable Organic Carbon Determination Using Flow-Cytometric Enumeration and a Natural Microbial Consortium as Inoculum FREDERIK A. HAMMES AND THOMAS EGLI* Department of Environmental Microbiology, Swiss Federal Institute for Environmental Science and Technology (EAWAG), U ¨ berlandstrasse 133, CH-8600 Du ¨ bendorf, Switzerland

The concentration of easily assimilable organic carbon (AOC) largely determines the microbiological stability of drinking water. However, AOC determination is often neglected in practice due to the complex and tedious nature of the conventional bioassay. The three major drawbacks of the conventional method are (1) a long assay time of 9-12 days, (2) the use of a labor-intensive enumeration technique (plating on growth media), and (3) limited information supplied by the use of selected pure cultures (Pseudomonas fluorescens P-17 and Spirillum NOX) for measuring a complex pool of natural bioavailable carbon compounds. A new method is proposed here, in which plating was replaced with fluorescence staining of total nucleic acids combined with flow cytometry as a rapid and straightforward growth enumeration method. This approach also allowed for the detection of inactive and/or unculturable microorganisms. Hence, the conventionally used pure cultures were replaced in the new AOC assay with a natural microbial consortium. It was shown that the flow-cytometric enumeration method could be used to establish complete growth curves for a natural microbial consortium growing on AOC. Compared to the end-point measurements of the conventional method, such kinetic data provide much clearer insight into the actual growth potential of a water.

Introduction Assimilable organic carbon (AOC) is a collective term describing the fraction of labile dissolved organic carbon that is readily assimilated by microorganisms, resulting in growth. It consists of a broad range of low molecular weight organic carbon molecules such as sugars, organic acids, and amino acids. AOC is a critical parameter for drinking water treatment and distribution processes. It represents only a small fraction (0.1-9%) of the total organic carbon (TOC) in water, but it is regarded as one of the main factors governing heterotrophic growth, and thus biological water stability (1, 2). Several studies have linked AOC directly to microbial regrowth and biofilm formation in drinking water distribution systems (2, 3). In addition, the present trend in drinking water treatment is toward a combination of treatments which will * Corresponding author phone: +41-44-823 51 58; fax: +41-44-823 55 47; e-mail: [email protected]. 10.1021/es048277c CCC: $30.25 Published on Web 03/31/2005

 2005 American Chemical Society

completely remove the nutrients that support microbial growth, namely, carbon (C), nitrogen (N), and phosphorus (P) (1). Based on the C:N:P molecular ratio of bacteria and biomass (100:10:1), the growth determining factor in most waters is usually carbon, i.e., AOC (1). These points emphasize the need for a reliable, realistic, and easily applicable AOC determination method. Conventional AOC analysis is done with a bioassay that was originally developed by Van der Kooij and co-workers (4) and later adapted by others (1, 5, 6), of which a detailed description is presented in Standard Methods (7). In short, this bioassay quantifies, by plating, bacterial batch growth as the number of colony forming units (cfu) in a water sample from inoculation until stationary phase, usually performed as an “end-point” measurement only. Pure cultures of either Pseudomonas fluorescens (P. fluorescens) strain P-17 or Spirillum strain NOX are most often used as test organisms, due to their different nutritional capabilities. After inoculation (prescribed at 500 cfu mL-1), the water sample is incubated at 15 °C for 9 days, and microbial growth is quantified on days 7, 8, and 9 with plating on nutrient agar. The result (average net growth) is thereafter related to the growth of the test organisms on pure solutions of acetate (P-17) or oxalate (NOX) by means of prederived yield values, and the final result is given as acetate-C equivalents (7). Typically, a water sample is considered to be biologically stable if it has an AOC concentration of less than 10 µg L-1 acetate-C equivalents, though this value depends on the presence of residual chlorine in the system (3), and whether indeed the water is limited in carbon and not phosphorus (8). Some improvements on the original method have been proposed by Kaplan et al. (5) and LeChevallier et al. (6) and included the use of a higher inoculum density ((2-4) × 104 cfu mL-1), a higher incubation temperature (specifically 22 °C for P-17). LeChevallier et al. (6) also proposed adenosine triphosphate (ATP) analysis instead for an estimation of bacterial growth, but plating remains the most generally applied enumeration method. Despite the relevance of AOC data to the industry, AOC measurement is still not performed routinely in practice and/or used often in the design and optimization of drinking water treatment, storage, and distribution systems. We believe that this is because the conventional method is somewhat tedious, time-consuming, and labor-intensive. Moreover, we question to which extent a pure culture(s), preconditioned on a single pure AOC compound, is able to give a reliable reflection of the AOC content of complex natural waters. Flow cytometry coupled with fluorescent staining has emerged as the leading tool for single-cell analysis in microbiology. Multiple examples already exist of the application of flow cytometry for the enumeration of either total bacteria or specific bacterial groups using either total nucleic acid or selective RNA stains (9). Probably the biggest advantage of flow cytometry is that it allows for rapid and accurate enumeration of all cells, including those which are unculturable or inactive. An obvious offspring of this is the ability to quantify a natural microbial consortium consisting largely of microbial cells that cannot be cultured on conventional media. In this study we describe the use of a recently patented method that entails fluorescent staining and flow cytometry instead of conventional plating or ATP analysis for quantification of microbial growth during the AOC measurement (10). This method has the advantage that it is fast, reliable, and reproducible. It is furthermore demonstrated that when combined with a natural microbial consortium and kinetic growth measurements, a more realistic interpretation of AOC and of the microbial regrowth VOL. 39, NO. 9, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3289

potential in natural water samples is obtained as compared with only end-point data.

Experimental Section Bacterial Strains and Natural Microbial Consortium. A pure culture of P. fluorescens strain P-17 (ATCC 49642) (from here on referred to as P-17) was obtained from D. Van der Kooij (Kiwa, Nieuwegein, The Netherlands). The culture was stored in 20% glycerol at -70 °C and cultivated on R2A agar (Oxoid) (15 °C, 3-5 days) prior to use. AOC-free stock solutions of the pure culture were prepared as described in Greenberg et al. (7), and these solutions were stored for up to 3 months at 5 °C. An AOC-free solution of a natural microbial consortium was prepared as follow: 40 mL of water, sampled directly after an activated carbon filter unit in a drinking water treatment plant, was filtered (0.22 µm, Millex-GP, Millipore) for removal of particulate organic carbon, inoculated with 100 µL of unfiltered water, and incubated in AOCfree vials without further amendments at 30 °C for 14 days. The cells were subsequently harvested by centrifugation (10 min, 3000g) and resuspended in HPLC-grade water (Fluka, CH) amended with a mineral buffer as described by LeChevallier et al. (6). This solution was then incubated for a further 7 days to ensure that all residual organic carbon had been degraded. Necessary controls were performed to validate the AOC-free status of the inoculum solutions. Preparation of AOC-Free Materials. Borosilicate glass vials (40 mL) with screw caps containing TFE-lined silicone septa were used for the assays. Carbon-free vials were prepared as described in Greenberg et al. (7) as well as in Charnock and Kjønnø (11). In short, the vials and screw caps were first washed with common detergent and thereafter rinsed thrice with deionized water. These were then submerged overnight in 0.2 N HCl and again rinsed thrice with deionized water. The vials were subsequently heated in a Muffel oven to 550 °C for at least 6 h (to remove all trace organics). The screw caps were soaked in a 10% sodium persulfate solution at 60 °C for at least 1 h, rinsed twice with deionized water and once with HPLC-grade water and finally air-dried. Growth of P-17 on Acetate-C. A pure acetate solution containing 100 µg L-1 acetate-C (added as sodium acetate; Fluka; technical grade) was prepared with bottled mineral water. Bottled mineral water was specifically used because it is well-buffered and contains all the required minerals for microbial growth. Aliquots (40 mL) of this acetate solution were then filtered with non-AOC-releasing filters (0.22 µm) into AOC-free vials, which were capped tightly and then pasteurized (70 °C, 30 min). Control samples were prepared in exactly the same way except that no acetate was added. For the non-AOC-releasing filters, we have used Millex-GP (Millipore) syringe-mounted filters which were flushed with ultrapure water before use. Triplicate vials were inoculated with 100 µL of the AOC-free P-17 inoculum, giving a final concentration of approximately 1 × 104 ((4 × 102) cells mL-1 (measured with flow cytometry) or 8 × 103 ((8 × 102) cfu mL-1 (measured with plating), and incubated at 15 °C for 10 days. Before sampling, vials were shaken for 30 s and thereafter 1 mL aliquots were removed with a sterile pipet for both flow cytometry and plating. Samples were taken at regular intervals and analyzed by means of plating and flow cytometry. For plating, samples were taken on designated sampling days, serially diluted in physiological salt solution (8.5 g L-1 NaCl), and plated on R2A agar using the spreadplate method. The plates were incubated at 15 °C for 3-5 days before counting. Growth of a Natural Microbial Consortium on AcetateC. An acetate solution containing 100 µg L-1 acetate-C and an acetate-free control solution were prepared as described above. Aliquots (40 mL) of these solutions were then filtered 3290

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 9, 2005

into AOC-free vials with a syringe-mounted filter (0.22 µm). The vials were not pasteurized again, as this step was deemed unnecessary when a natural microbial consortium is used as inoculum. Triplicate vials were inoculated with 100 µL of the AOC-free consortium, giving a final concentration of 5 × 103 ((2 × 102) cells mL-1 (measured with flow cytometry) and incubated at respectively 15 and 30 °C for 9 days. Samples were taken at regular intervals and analyzed by means of flow cytometry. Kinetic Measurements with the Natural Consortium. For a first kinetic experiment, an acetate-C calibration series was prepared similar to the acetate solutions above, using commercial mineral water as matrix and control, at the following concentrations: 0 (control), 50, and 100 µg L-1 acetate-C. In a second experiment, natural river water (Chriesbach river, Du ¨ bendorf, CH) was filtered (0.22 µm) and diluted with commercial mineral water to final concentrations of 1, 5, 10, 50, and 100% (undiluted) river water. Water samples from both experiments were filtered and inoculated with the natural consortium as described above and incubated at 30 °C. Samples were taken at t ) 0, 12, 14, 16, 18, 20, 22, 26, 48, and 52 h after inoculation for the acetate experiment and at t ) 0, 8, 10, 12, 14, 16, 19, 21, and 23 h for the river water experiment. Samples were stained and analyzed with flow cytometry as described below. The specific growth rates (µ) for the natural microbial consortium in each sample were determined as follows:

µ ) (ln(x) - ln(x0))/∆t

(1)

where x, x0 are the microbial concentration measured at two time points and ∆t is the expired time interval between these points. Fluorescent Staining and Flow Cytometry. All enumeration samples (1 mL) were amended with 1% lysis buffer (10% Triton X-100, 5% Tween 20, 10 mM TrisHCl, and 1 mM ethylenediaminetetraacetic acid (EDTA)) and directly stained with 10 µg mL-1 SYBR Green stain (1:100 dilution in dimethyl sulfoxide (DMSO); Molecular Probes), and incubated in the dark for at least 20 min before measurement. Where necessary, samples were diluted after staining in filtered (0.22 µm) mineral water, so that the concentration measured in the flow cytometer was always less than 3 × 105 counts mL-1. Flow cytometry was performed using a PASIII flow cytometer (Partec, Mu ¨ nster, Germany) equipped with a 25 mW argon ion laser, emitting at a fixed wavelength of 488 nm. Green fluorescence was collected in the FL1 channel (530 ( 30 nm) and red fluorescence collected in the FL3 channel (>590 nm). All parameters were collected as logarithmic signals. Data were analyzed using Flowmax software (Partec). The specific instrumental gain settings were as follow: FL1 ) 420, FL3 ) 700, speed ) 3 (implying a count rate of less than 500 events s-1). All samples were triggered on green fluorescence (FL1). Staining results of selected samples were controlled with fluorescence microscopy, using an inverted microscope (Olympus IX 51) equipped with the appropriate filters.

Results Flow-Cytometric Absolute Cell Counting. SYBR green stains total nucleic acids and emits, upon excitation at 488 nm, a bright fluorescent signal which was detected on the FL1 channel at 530 ( 30 nm. Due to the emission spectrum of the stain (see www.probes.com for details), a corresponding signal was also detected above 590 nm (FL3 channel). Parts A-F of Figure 1 show examples of typical histogram and dot plot data generated when a sample of P. fluorescens strain P-17 (Figure 1A-C) and a natural microbial consortium (Figure 1D-F) is analyzed with flow cytometry. Gates were defined on the combined dot plot (Figure 1C,F), to discrimi-

FIGURE 1. Typical examples of fluorescence histogram data and enumeration gates for P. fluorescens strain P-17 in pure water (A-C) and a mixed microbial consortium in river water (D-F), after staining with SYBR green and flow-cytometric analysis. The x-axis shows green (FL1) or red (FL3) fluorescence intensity, and the y-axis shows the number of events recorded for a corresponding fluorescence intensity. Gates R1 on the combined fluorescence dot blots (C and F, respectively) were used for enumeration purposes, allowing clear distinction between the background and the stained sample in natural water. nate between positive signals and background using both green and red fluorescence intensity of the particles. Note that the intensity of the background signal tended to vary depending on the water used. While deionized laboratory water resulted in practically no visible background (Figure 1C), some natural water samples displayed severe background, most probably as a result of mineral formation in calcium-rich water (Figure 1F). Hence, for the sake of standardization, the gating on the dot plot was used in all cases. The enumeration method was validated with a stock solution of P-17. This control revealed the standard error of the enumeration method to be 4.8%, of which 1.4% could be ascribed to machine error and the rest to error during sampling, dilution, and staining. Moreover, the counts obtained with flow cytometry were confirmed for P-17 with conventional plate counts on R2A agar, and the plating efficiency exceeded 90% (data not shown). Flow-Cytometric Enumeration Compared to Conventional Plating. Flow cytometry coupled with fluorescent staining is a simple, fast, and statistically reliable method for bacterial enumeration. Figure 2 shows a comparison between this method and conventional plate counting in an AOC assay of 100 µg L-1 acetate-C, inoculated with P-17 and incubated at 15 °C. It is significant to note that in the crucial stationary phase period of the assay (ca. day 5 onward), the flow cytometry data displayed much more stability than the plating data. This could be ascribed to the plating method being prone either to human error or to physiological changes to the bacteria (e.g., loss of activity and/or culturability). In terms of statistical accuracy, the plate count method can reliably detect between 30 and 300 events/plate, which are converted to the final result based on the appropriate dilution factor. For example, a sample containing 3 × 105 cfu mL-1 would require at least two 10-fold dilutions. Flow cytometry

detects up to 6 × 104 events per analysis (the total volume analyzed is 200 µL), which means that a sample containing 3 × 105 cells mL-1 can be analyzed without dilution. The average net growth in the samples was determined as the difference between the average concentration on days 9 and 10 and the concentration directly after inoculation (t ) 0). The control values were then subtracted from the acetatecontaining samples. On the basis of the empirical yield values of 4.1 × 106 cfu (µg of acetate-C)-1 (7), the net growth recorded in Figure 2 yields AOC values of respectively 137 ( 26 µg L-1 acetate-C for the plating method and 124 ( 9 µg L-1 acetate-C for the flow-cytometry method. Evidently, the expected AOC concentration was slightly overestimated in both instances. It may well be that the pasteurization step used in these analyses resulted in some minor carbon contamination of the vials. The specific growth rate (µ) with acetate was approximately 0.07 h-1, and in the control it was approximately 0.01 h-1. Natural Microbial Consortium and Temperature. A combination of fluorescent staining and flow cytometry enabled the enumeration of a natural microbial consortium, with flow cytometry data similar to those presented in Figure 1D-F. This method was used to assess the effect of different incubation temperatures on the AOC assay with pure acetate solutions. Figure 3 shows the consortium proliferating much more rapidly with acetate at 30 °C than at 15 °C. Stationary phase was reached at 30 °C between 30 and 40 h after inoculation in both the control and the acetate-containing sample. In the samples inoculated at 15 °C, stationary phase was only reached after 70-130 h of incubation. The maximum specific growth rates at 30 °C were approximately 0.18 and 0.14 h-1 for the acetate-containing and control samples, respectively, and at 15 °C it was 0.09 and 0.07 h-1, respectively. Moreover, the empirical yield values of 6.08 × 106 cells VOL. 39, NO. 9, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3291

FIGURE 4. Batch growth curves for a natural microbial consortium growing with acetate during 52 h of incubation at 30 °C (where 9e+5 equals, for example, 9 × 105). The maximum specific growth rates in these samples were respectively 0.199 h-1 (r2 ) 0.992; b), 0.210 h-1 (r2 ) 0.997; O), and 0.202 h-1 (r2 ) 0.999; 1). Error bars indicate standard deviation on triplicate samples.

FIGURE 2. Batch growth curves for P. fluorescens P-17 growing with and without acetate as determined with plating (A) and flow cytometry (B), respectively (where 9e+5 equals, for example, 9 × 105). Error bars indicate standard deviation on triplicate samples.

FIGURE 3. Batch growth curves for a natural microbial consortium growing with and without acetate at 15 and 30 °C, respectively (where 1e+6 equals, for example, 1 × 106). Error bars indicate standard deviation on triplicate samples. (µg of acetate-C)-1 at 30 °C and 4.96 × 106 cells (µg of acetateC)-1 at 15 °C were higher than the values given in Standard Methods (4.1 × 106 cfu (µg of acetate-C)-1) (7) but in a similar range to those recorded for P-17 growth on acetate in this work (5.1 × 106 cells (µg of acetate-C)-1) (Figure 2). Temperature also affected the yield measured as cell numbers. This was evidenced by the yield in cell numbers at 30 °C, which was notably higher than that recorded at 15 °C. The significantly higher specific growth rate at 30 °C, which resulted in stationary phase being reached more rapidly, was the reason this incubation temperature was favored in further experiments. 3292

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 9, 2005

Kinetic Analysis of Growth on AOC. Batch growth curves of the natural microbial consortium on a single AOC compound (acetate) at two concentrations were considerably different from batch growth curves on a natural AOC mixture (river water). Exponential growth was detected between 12 and 26 h after inoculation in the acetate-containing solutions (Figure 4), with no significant differences in the specific growth rates recorded for the various concentrations (0.199-0.202 h-1; Figure 4). Note that the end-point cell concentrations (t ) 48 and 52 h) correspond perfectly to the added acetate concentrations. When the average cell concentration of the control was subtracted from the average values for the acetate-containing samples, the resulting cell concentration in the 50 µg L-1 acetate-C sample was 49.8% of the concentration in the 100 µg L-1 acetate-C sample. Independent triplicate vials that were only sampled after 72 h confirmed the end-point values, which also indicates that no significant carbon contamination occurred as a result of continuous sampling during the experiment. Figure 5A shows batch growth curves for the natural consortium on five dilutions of river water. Throughout the measuring period, the different concentrations could be distinguished from one another, and exponential growth was recorded in all samples from as early as 8 h after inoculation (Figure 5B). In fact, contrary to the acetate result above, it was apparent that different natural AOC concentrations resulted in different specific growth rates (Figure 5B). The lowest detected specific growth rate was 0.164 h-1 for the 1% dilution, and the highest detected specific growth rate was 0.351 h-1 for the 100% river water sample. End-point measurements predicted less accurately the AOC concentration than when acetate was used as carbon source, especially at higher dilutions of the river water. This can be ascribed to the influence of the dilution water, e.g., a different pH, additional carbon or additional micronutrients. Alternatively, it is conceivable that different growth conditions might have affected the final cell size. Thus, growth measured only as cell numbers might err slightly in the estimation of total biomass increase.

Discussion The conventional AOC bioassay is not used as commonly in practice as one would expect from its importance for drinking water production. This apparent neglect is in our opinion due to the tedious nature of the conventional method and also because of inevitable questions arising from the use of a pure culture(s) to accurately assess a wide range of natural AOC components. The aim of this work was to examine fluorescence staining of total microbial nucleic acids coupled

FIGURE 5. Batch growth curves for a natural microbial consortium growing on five dilutions of river water (A) and calculation of the maximum specific growth rates in these samples (B), which were respectively 0.16 h-1 (r2 ) 0.994; b), 0.24 h-1 (r2 ) 0.995; O), 0.25 h-1 (r2 ) 0.991; 1), 0.3 h-1 (r2 ) 0.995; 3), and 0.35 h-1 (r2 ) 0.998; 9) (where 6e+5 equals, for example, 6 × 105). Error bars indicate standard deviation on triplicate samples. with flow-cytometric quantification and a natural microbial consortium as a means of improving the AOC bioassay. In this study (e.g., Figure 2) it was shown that flow cytometry can be used in combination with the conventional bioassay (P-17, 15 °C) instead of other enumeration methods. While plating is relatively easy and does not require specific equipment, flow cytometry is more reproducible and less susceptible to variations in culturability of the cells, and, as an additional benefit, the cell count is obtained within 15 min (10 min staining, 5 min analysis) instead of after 1-3 days, as are required for the incubation of the plates. Total nucleic acid fluorescence staining and flow cytometry also allows the detection and enumeration of all bacteria in a sample, including those that are inactive or unculturable. This means that growth is correctly assessed as all organisms that proliferated in the sample during the assay and not only those that are both viable and culturable at the sampling end points. For example, large fluctuations in plate count results were reported by LeChevallier et al. (6) when P-17 was grown on acetate at 30 °C. In fact, using acridine orange direct microscopic counts, Kaplan et al. (5) demonstrated that only between 60 and 80% of the cells in pure-culture AOC assays are culturable and that this might affect the determined AOC value when predetermined yield values are used. Compared to ATP analysis, flow cytometry is more sensitive and less prone to background interference, and it analyses specifically the numerical growth of a culture, which is the fundamental principle of the AOC concept, rather than a cell-associated parameter (ATP). This is of particular significance since different bacteria can have widely different

intracellular ATP concentrations during the various stages of growth. For example, LeChevallier et al. (6) reported nearly 10-fold differences in mean ATP values of P-17 (1.85 fg cell-1) and NOX (0.213 fg cell-1), respectively. Moreover, the use of ATP analysis requires two separate conversions to achieve AOC data: a first conversion from total ATP to total cell numbers and a second conversion from total cell numbers to AOC. The use of a pure culture for detection of natural AOC is contentious as such. Indeed, although P-17 was initially recognized and used as a typical heterotrophic strain representing drinking water bacteria, Spirillum strain NOX was subsequently added to the AOC determination method as it was discovered that P-17 does not adequately detect ozonation products (1). Flavobacterium sp. strain 12 has also been suggested for specific waters (targeting sugars in particular), while several other natural isolates have also been tested (1, 11). However, because AOC covers a wide range of assimilable substrates including for example organic acids, sugars, alcohols, amino acids, and oligopeptides in various molecular sizes, a pure culture will probably never suffice to detect all of these (12). Using combinations of these pure cultures as a mixed inoculum has been examined but presents experimental problems in terms of both the enumeration and interpretation of the results (1). As a result, researchers often prefer to use only one of these strains (5). Even though it was not specifically tested in this work, it is possible to measure with flow cytometry analysis any other pure culture or combinations of two or more pure cultures in the AOC bioassay. In the latter case it should be noted that the total nucleic acid staining method described in this study would suffice only for the detection of the total number of grown cells. However, application of fluorescent in situ hybridization (FISH) probes and fluorescent antibodies together with flow cytometry would enable the separation and enumeration of the individual subpopulations of a natural consortium (9). Though the latter option may not yet be lucrative for industrial applications, it does offer immense research opportunities in the field of microbial ecology, e.g., following microbial population dynamics in the presence of specific AOC compounds. An alternative approach to pure cultures is to use a natural microbial consortium. Such a consortium possesses a much broader and diverse substrate range than a single pure culture and should thereby inherently be able to offer a more realistic interpretation of the actual AOC content when natural substrates are assayed. This approach adds the possibility to use an in situ natural microbial consortium in the AOC analysis, in other words, a natural microbial consortium autochthonous to the water being analyzed and, therefore, adapted to the types of AOC in such a water sample. Flow cytometry is a straightforward method whereby a natural microbial consortium can be enumerated accurately, obviating the problem of nonculturability. This work demonstrated that when a natural microbial consortium, which was adapted to 30 °C, was used in the assay, stationary phase end-point data could be achieved within 30-40 h following inoculation. As such, this is significantly faster than any other previously reported method (6). It remains to be tested extensively the degree to which different natural microbial consortia would produce similar AOC results on the same water. Preliminary experiments in this regard with four inoculums produced from different sources showed no statistical difference in the predicted AOC value (data not shown). Using a natural microbial consortium, however, poses a question as to how microbial growth should best be related to carbon equivalents, because standardizing the yield of a natural consortium on natural AOC entails a rather complicated argument. Empirical yield values for pure culture VOL. 39, NO. 9, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3293

growth on acetate are used in the conventional AOC method, but these are subject to several factors. First, these yield factors are dependent on temperature (15 °C in the conventional method) and require recalculation if a different incubation temperature is used (6, 7). Second, the yield values differ significantly for different microbial strains when used under otherwise similar conditions. The best examples of this are P-17 and NOX, which have yield values of 4.1 × 106 and 1.2 × 107 cfu (µg of acetate-C)-1 respectively. Third, the yield values will also differ significantly when different pure carbon sources are used. For example, NOX yields 1.2 × 107 cfu (µg of C)-1 on acetate, while only yielding 2.9 × 106 cfu (µg of C)-1 on oxalate (1). Kaplan et al. (5) illustrated with DOC mass balances during AOC measurements that P-17 yield on acetate is about 1.25 times lower than its yield during growth on complex, natural substrates. In fact, Van der Kooij (1) reported various yield values for five different pure cultures growing on various carbon sources (acetate, oxalate, glucose, starch, lactate) ranging between 2.9 × 106 and 1.2 × 107 cfu (µg of C)-1. This latter point is even further complicated by the use of empirical yield values derived from growing a preconditioned pure culture on a single pure carbon source, to convert growth on natural complex AOC to carbon equivalents. Alternative to the above experimental approach, a theoretical argument could have its merit. Batte et al. (13) suggested the average carbon content of a bacterial cell to be 2.0 × 10-14 g. From this value it can be deduced that 1 µg of carbon equals 5.0 × 107 cells. Therefore, if 1 µg of AOC is utilized, and for example 50% thereof is actually assimilated, then the theoretical yield would be 2.5 × 107 cfu (µg of AOC)-1. However, also this argument is fundamentally flawed, since microbial cells sizes and carbon content differ tremendously, as does the percentage of carbon being assimilated by the cells (14). Hence, for the sake of simplicity, we recommend the use of a constant value of 1 × 106 bacteria (µg of AOC)-1 as suggested by Van der Kooij (1) for the conversion of natural microbial consortium growth to AOC values, representing a value close to the maximum yield of bacteria on AOC (1). An up to now untested argument which needs to be considered is that cellular size might vary for different bacteria grown on different AOC substrates. The implication thereof is that total yield should be a function of both the cell numbers and the cell size. However, first flow cytometric measurements of cellular size of the natural microbial consortium used in this study have revealed no significant differences in cell size between the inoculum and those cells that have entered the stationary phase after all AOC was consumed, irrespective of the water that was tested (data not shown). The proposed AOC method could easily be automated, and for this purpose it has been patented (10). Yet, even without automation, the simplicity with which the flowcytometry-based method can measure a large number of samples enabled the recording of batch growth curves as depicted in Figures 4 and 5A. This approach has considerable additional benefits over the only end-point measurements of the conventional method, as it provides a more complete and realistic interpretation of microbial growth on AOC. For example, growth inhibitory substances, AOC quality (type), and AOC quantity all have profound effects on the growth kinetics. With regard to the latter, it was initially explored whether the kinetic approach could be used as a tool to predict the AOC concentration of a sample, based on the

3294

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 9, 2005

specific growth rate, even before the stationary phase is reached. This notion is supported to some extent by the results depicted in Figure 5A,B for growth on natural AOC. It has, however, to be taken into account that numerous factors apart from the carbon concentration could influence the specific growth rate of a natural microbial population in a water sample from the environment or from a technical system. This kinetic approach adds value not only to the field of AOC measurement but also to the general study of natural microbial growth under carbon-limited conditions.

Acknowledgments We appreciate the critical evaluation of the manuscript by Dr. Martin Rieger and technical assistance from Verena Loser.

Literature Cited (1) Van der Kooij, D. Assimilable Organic Carbon (AOC) in Treated Water: Determination and Significance. In Encyclopedia of Environmental Microbiology; Bitton, G., Ed.; John Wiley & Sons: Hoboken, NJ, 2002; pp 312-327. (2) Escobar, I. C.; Randall, A. A.; Taylor, J. S. Bacterial Growth in Distribution Systems: Effect of Assimilable Organic Carbon and Biodegradable Dissolved Organic Carbon (BDOC). Environ. Sci. Technol. 2001, 35, 3442-3447. (3) Van der Kooij, D. Assimilable Organic Carbon as Indicator of Bacterial Regrowth. J. Am. Water Works Assoc. 1992, 84, 57. (4) Van der Kooij, D.; Visser, A.; Hijnen, W. A. M. Determination of Easily Assimilable Organic Carbon in Drinking Water. J. Am. Water Works Assoc. 1982, 74, 540-545. (5) Kaplan, L. A.; Bott, T. L.; Reasoner, D. J. Evaluation and Simplification of the Assimilable Organic Carbon Nutrient Bioassay for Bacteria Growth in Drinking Water. Appl. Environ. Microb. 1993, 59, 1532-1539. (6) LeChevallier, M. W.; Shaw, N. E.; Kaplan, L. A.; Bott, T. L. Development of a Rapid Assimilable Organic Carbon Method for Water. Appl. Environ. Microb. 1993, 29, 1526-1531. (7) Greenberg, A. E., Clesceri, L. S., Eaton, A. D., Eds. Standard Methods for the Examination of Water and Wastewater, 18th ed.; American Public Health Association: Washington, D.C., 1993. (8) Lehtola, M. J.; Miettinen. I. T.; Vartiainen, T.; Myllykangas, T.; Martikainen, P. J. Microbially Available Organic Carbon, Phosphorus, and Microbial Growth in Ozonated Drinking Water. Water Res. 2001, 35, 1635-1640. (9) Porter, J.; Pickup, R. W. Nucleic Acid-Based Fluorescent Probes in Microbial Ecology: Application of Flow Cytometry. J. Microbiol. Methods 2000, 42, 75-79. (10) Hammes, F.; Egli, T. A Method for Assimilable Organic Carbon Determination and Microbial Growth Inhibition Assessment in Water That Can Be Automated. EU Pat. Appl. 04025113.4, 2004. (11) Charnock, C.; Kjønnø, O. Assimilable Organic Carbon and Biodegradable Dissolved Organic Carbon in Norwegian Raw and Drinking Waters. Water Res. 2000, 34, 2629-2642. (12) Hem, L. J.; Efraimsen, H. Assimilable Organic Carbon in Molecular Weight Fractions of Natural Organic Matter. Water Res. 2001, 35, 1106-1110. (13) Batte, M.; Koudjonou, B.; Laurent, P.; Mathieu, L.; Coallier, J.; Prevost, M. Biofilm Responses to Aging and to a High Phosphate Load in a Bench-Scale Drinking Water System. Water Res. 2003, 37, 1351-1361. (14) Linton, J. D.; Stephenson, R. J. A Preliminary Study on Growth Yields in Relation to the Carbon and Energy Content of Various Organic Growth Substrates. FEMS Microbiol. Lett. 1978, 3, 9598.

Received for review November 5, 2004. Revised manuscript received February 1, 2005. Accepted February 14, 2005. ES048277C